PROJECT SUMMARY All animals, from worms to humans, distinguish between sensory information generated by its own movements (i.e., reafference) and that produced by external stimuli (i.e., exafference). However, because the same sensory receptors detect reafference and exafference, accurate segregation of sensory information is not a trivial computation. Theoretical neuroscientists have hypothesized that animals must compute a representation, or internal model, of the sensory consequences of a given self-generated movement. Then, with each motor command, a motor copy is conveyed to a neural structure serving as a comparator, where it can be compared with the actual reafference arriving from the periphery. By removing expected reafference, the unexpected component remains and can direct attention toward potentially important stimuli. For infant animals, the computational difficulty of this task is further compounded by the changes caused by normal growth. This continual change necessitates that the internal model be continually updated. Thus, infant animals must be able to update their internal models even as their neural systems are developing. In a preliminary experiment, we found that rats at 8 days of age are able to rapidly integrate the unexpected reafference arising from a weight being attached to the forelimb. Specific Aim 1 of this proposal tests whether neural inputs to the external cuneate nucleus (ECN), a brainstem nucleus that processes sensory input from the forelimb and acts as a comparator of expected and actual reafference, mediates this updating process. Updating a representation of expected reafference requires a reliable signal of actual reafference, so that a deviation from expectations can be assessed. Since the ECN gates expected reafference from self-generated wake movements, it is unclear how downstream structures receive a reliable signal of actual reafference. However, gating of reafference in the ECN occurs in a state-dependent manner, being engaged during wake but disengaged during sleep. Thus twitches, which are self-generated movements that occur exclusively and abundantly during sleep, reliably activate sensorimotor circuits across the neuraxis. This presents two types of self-generated movement that could contribute to updating expected reafference; wake movements with unexpected reafference and twitches. Therefore, Specific Aim 2 tests whether sleep- related twitches contribute to this updating process. Finally, previous work in the cerebellum of adults has supported its role in computing internal models for comparing expected and actual reafference. Although the adult-like cellular circuitry of the cerebellum is not established until after the second postnatal week in rats, many of the key components are functionally active earlier in development. Specific Aim 3 assesses, for the first time in early infancy, whether the cerebellum contributes to the updating of expected reafference. All together, these experiments will provide critical new information about the behaviors and neural processes that contribute to the earliest abilities of infant animals to distinguish between self and other.